Skip Navigation

What Is Optimization-as-a-Service?

Resources

41% of business decision makers worldwide agree that digital experimentation will be the top driver of growth in the next three years — that’s higher than AI, automation, and mobile.

The future of digital experimentation is bright. It also has a new name: Optimization-as-a-Service.

Optimization-as-a-Service is the latest offering from Optimizely, a Leader in the Gartner Magic Quadrant for Digital Experience Platforms. The solution takes the first-in-class digital experimentation capabilities of Optimizely — complete with A/B and multivariate testing — and turns them on hyperdrive, thanks to personalized behavior targeting and 1:1 content and product recommendations.

Whether you’re looking to increase AOVs, revenue per customer, lead submissions, or engagement, Optimization-as-a-Service can help you meet — and surpass — your conversion goals. Just take a look at what Optimization-as-a-Service did for a few companies you may recognize:

  • The Wall Street Journal saw a 64% lift in conversion rate for digital subscriptions

  • HP enjoyed a $21 million revenue increase over 500 campaigns.

  • Crate & Barrel increased revenue per visitor by 20%.

These aren’t isolated incidents. After implementing experimentation, 75% of companies experience at least a 5% increase in revenue — solely attributable to the solution. Almost half of companies experience a 10% increase in revenue or more.

Interested? Then read on. We break down everything you need to know about Optimization-as-a-Service, its key features, and how to get started. 


What Is Optimization-as-a-Service?

Launched in early 2021, Optimization-as-a-Service offers behavior targeting, A/B and multivariate testing, and personalized recommendations in a single, cloud-based solution. 

What makes Optimization-as-a-Service different? For starters, a company focused on digital experience typically would have to purchase three separate solutions to pursue their optimization strategy. The Forrester Wave on Experience Optimization Platforms defines these focus areas as behavioral targeting, online testing and experimentation, and recommendations:

Organizations would purchase an experimentation and testing software, one or more third-party targeting solutions, and a recommendation engine. None of these would speak to each other, siloing the data and insights and forcing marketers to connect the dots. 

Optimization-as-a-Service is the first and only solution to bring all three solutions under one roof. At last, marketers can enjoy a purpose-built recommendation engine, a world-class experimentation platform, and comprehensive targeting that all works together in real-time. 

Plus, there’s no need to replatform, so you can get up and running fast. Optimization-as-a-Service can be deployed on top of any CMS or Commerce Engine out there, from Sitecore to WordPress, thanks to a straightforward, client-side JavaScript solution.


Optimization-as-a-Service: The Features You Need to Know

Optimization-as-a-Service offers targeting, testing, and recommendations in one powerful solution. Let’s dig into these capabilities in more detail.

Behavioral Targeting

Targeting identifies behavioral audiences among your visitors, enabling you to provide more personalized experiences. Based on targeting, visitors may be presented with entirely different landing pages, or simply with different images or content that are more relevant to them. 

Key targeting features of Optimization-as-a-Service include:

  • Behavioral targeting based on attributes like location, new vs. returning visitors, language, device, and UTM parameters

  • More sophisticated targeting via third-party integrations (Optimization-as-a-Service offers one-click integrations with popular customer data platform (CDP) and data management platform (DMP) solutions like Krux, Lytics, Google Ads, Tealium, Lotame, and Demandbase)

  • AI-based segmentation that uses artificial intelligence to help you discover opportunities for new audiences, based on real-time interactions that categorize visitors into interest-based segments


A/B and Multivariate Testing

Optimization-as-a-Service includes both A/B and multivariate testing. With A/B testing, you split your traffic into two groups and test a single variable, such as a page layout or a button color. When the results declare a winner, you apply the winning variation to 100% of your audience.

With multivariate testing, you can test combinations of variables at the same time. For example, if you’re testing a new lead form, you can test 3 form fields against 4, at the same time you test different call-to-action buttons..  

With a browser-based WYSIWYG editor, Optimization-as-a-Service lets you quickly design new tests. You can swap out text, images, layout, HTML, and other front-end changes with ease. Your UX designers are going to love this.

We saved our favorite feature for last: multi-armed bandits. Powered by machine learning algorithms, the solution automatically allocates more traffic to the winning variation to help you reap the rewards even faster, while you’re still running the test.

Recommendations Engine

Testing helps you determine what performs better, according to the masses. But what about each individual within that mass? What is the single most relevant thing you could show a particular visitor to get them to convert, based on their real-time behavior and AI and machine learning models?

That’s where the Optimization-as-a-Service recommendations engine comes in. While testing occurs at a cohort level, this performs on an individual level, providing 1:1 content and product recommendations to visitors. 

The engine’s content recommendations rely on natural language processing and artificial intelligence to deliver the “next best content” to a visitor, avoiding articles they’ve previously read and including ones that are most likely to drive the desired behavior. 

 

Production recommendations are based on 10+ out-of-the-box machine learning algorithms proven to increase AOV and order volumes.

Bringing It All Together

Any one of these capabilities is exciting on its own, but the true wow factor comes from the way they work together. Here’s a look at a hypothetical logic sequence in Optimization-as-a-Service targeting CMOs:

The targeted audience is CMOs. Visitors who meet this target audience will be shown different test variations. Within those variations, they’re also served with individualized content recommendations. The end result is a precisely targeted, highly personalized digital experience that’s unique to each visitor. 

Marketers can then review the analytics to see how it’s all working. Content analytics offer insights on how the current content is performing, and artificial intelligence pinpoints any missed opportunities. Experimentation analytics provides an overview of the testing results, and performance analytics show how individualized recommendations are driving ROI.

The analytics available with Optimization-as-a-Service are a game changer for marketers and become your competitive advantage. The days of basing your decisions on opinions and conventional “best practices” are over. Instead, you can use actual data to know what’s working, what’s not, and predict what will work better in the future.

Getting Started with Optimization-as-a-Service

Optimization-as-a-Service is available to Optimizely customers using Content Cloud and Commerce Cloud. It’s also available as a standalone solution that you can layer on top of your existing content or commerce platform, such as Sitecore, Drupal, Magento, or WordPress. That’s right — you don’t have to use everything Optimizely has to offer, you can just enjoy the benefits of this powerful targeting and testing solution.

That’s just one of the many things we love about Optimizely and their Optimization-as-a-Service offering. You can start optimizing in a matter of weeks, not months.

Whether you’re building a brand new site and want to start with optimization out of the gate, or you want to add it to an existing platform, we can help. Schedule a time to talk now.